eli_oat, renem, snarfed, [chrisaldrich] and [jeremycherfas] joined the channel
#[jeremycherfas]It seemed to me more of a no tracking thing. That if you follow a link from his site, your destination won’t know you came from there. That is how I read it.
[grantcodes] joined the channel
#[grantcodes]Is there any reason why micropub doesn't return the location of new posts / media as json? Trying to read headers is really killing me in this extension.
KartikPrabhu joined the channel
#Zegnat[grantcodes], I think the reason was so endpoint stuff is agnostic of this-years-favourite serialisation format. Authentication (access token) and success status (with location) are all using the HTTP protocol as their format.
#ZegnatBut not sure, as some stuff definitely seems pinned to JSON anyway.
#ZegnatNot sure how Known handles replies to replies, so can’t tell you.
#ZegnatI do find it interesting that your post is `in-reply-to` 4 separate URLs
#jeremycherfasWell, it is a comment on each of them. What should they be?
#ZegnatNot sure at all, and definitely not saying it is wrong, jeremycherfas!
#jeremycherfasFair enough. Manual till it hurts sometimes hurts too much.
#ZegnatI guess I am always a little inclined to make it a reply to only the one thing that made me write it. And not also a reply to things further up the reply chain. If that makes sense?
#ZegnatShouldn’t make a difference for the parsing end though
#dgold!tell [miklib] no-referrer is nothing to do with webmentions. It is solely so that if you follow a link on my site, the site you travel to does not know you've come from my site.
#jeremycherfasThat's what I thought, dgold, as you explained it on your page.
#jeremycherfas!tell zegnat I suppose I am trying to make sure that all the various posts tie together, but maybe that isn't necessary. People can follow the links if they choose to.
#ZegnatThere is definitely some value in your markup too jeremycherfas. I don’t see why consumers of the mf2 would have problems with it, so stick to it :D
#LoqiZegnat: jeremycherfas left you a message 1 minute ago: I suppose I am trying to make sure that all the various posts tie together, but maybe that isn't necessary. People can follow the links if they choose to.
#@petermolnar↩️ Webmention: one website to poke another website that an entry of mine (the first website) has your (the second website) link (URL), so the contents of the entry might be interesting to you.
The rest is up to the second website - process it? Display it? Etc. (twitter.com/_/status/998505847717318656)
#ZegnatOh, no, it does on an older version of curl but not on a more recent one
#Zegnatjgmac1106, I am not a protocol expert, but it looks to me as if your server advertises that it can do HTTP/2, but when curl tries to use it the server doesn’t respond correctly.
#ZegnatYou can add --http1.1 to the curl command so curl doesn’t upgrade, and that seems to work
#jgmac1106was hoping maybe I stumbled on to why micropub doesn’t work for me
#ZegnatIf the micropub tools try to correctly use HTTP/2 (which I doubt, actually, as a lot of tooling fails at HTTP/2) and are not getting anything back from your server, they would immediately fail
#ZegnatI wonder why browsers aren’t failing at your site though, as they usually try HTTP/2
#[jgmac1106]just would love to solve micropub, might try another 30 day free host to see if it my shared hosting provider
#petermolnar^^^ this thread is _really_ not related imo.
AngeloGladding joined the channel
#jeremycherfas!tell dgold Your nanopub functions array_replace_keys() decode_input() and recode_input() seem to be where I would look to make nanopub play nicely with Grav's YAML; is that correct?
snarfed, Kyle-K, Ruxton, leg, sebsel and [miklb] joined the channel; Kyle-K left the channel
#[miklb]!tell dgold that’s my point, without a webmention, using no-referrer means I will never know you linked to something I wrote. I realize they are two different technologies.
#Loqidgold: jeremycherfas left you a message 1 hour, 15 minutes ago: Your nanopub functions array_replace_keys() decode_input() and recode_input() seem to be where I would look to make nanopub play nicely with Grav's YAML; is that correct?
#Loqidgold: [miklb] left you a message 2 minutes ago: that’s my point, without a webmention, using no-referrer means I will never know you linked to something I wrote. I realize they are two different technologies.
[jgmac1106] joined the channel
#dgoldbut I'm not sure that the efficacy of you knowing that I wrote something about you supersedes the right of someone visiting my site to not be "tracked" on the internet
#dgoldjeremycherfas: that's precisely where you should be looking
#jeremycherfasCool. And that's where I am. Deep in the weeds.
#dgold[miklb]: in fact, I think that type of tracking is where I got onto this boat in the first place - if I wanted you to know I wrote about you, I'd tell you :-)
#[miklb]I mean, you could be writing something really bad about me or my writing, have a big audience, send them to my site, and without a webmention and using no-referrer I’d have to hunt to find where this was coming from. Who’s rights are being protected there?
#dgoldthe rights of the users on my site following a link
#dgoldwho I think should have the highest degree of protection
#[miklb]So I could in theory block all traffic coming from no-referrers then.
#ZegnatThere were whole no-referer-redirect-services out there. Lots of forums used those, so when you linked to something, they wouldn’t know you were coming from a (sometimes private) server. No loss to lose those pesky things for real.
#dgoldI can see a case for same-origin, to check the efficiency of your internal links, but I don't see a privacy-case supporting sowco
#jeremycherfasNoob php question: the frontmatter sent by Omnibear optionally contains an array called `tags`. Grav expects an element called `tag` to be nested under an element called `taxonomy`. I have tried `$frontmatter['taxonomy'] = $frontmatter['tags'];` and various versions, and nothing seems to work. Any tips gratefully received.
#jeremycherfasThaks petermolnar I have nearly solved it with `$frontmatter['taxonomy'][] = $frontmatter['tags'];`
#jeremycherfasJust need to fiddle a bit more.I think that may be it, kevinmarks
#[kevinmarks]that would do what you said, give you tag nested under taxonomy.
#[kevinmarks]though whether it wants an array or string you'll have to check
#jeremycherfasYup, that did it. Thanks very much. I got some of the way there. It'll take an array or a string, so I'm good.
#jeremycherfasAnd fortunately, probably, I have run out of time for further testing. But this is a good time to break.
#petermolnarbtw, a tip: dict (associative array) for tags, like $x['tagname'] = 1; array_keys() for the actual values => deduplication built in an very, very fast lookup for existing values
#[kevinmarks]though you usually don't have enough tags for that to br a problem, unless you're an instagram influencer
#[jgmac1106]I am marking up my course template. Would you give the course header an h-card as if it had identity or an h-entry?
[wiobyrne] joined the channel
#ZegnatI would only give persons (natural and legal) an h-card. So I wouldn’t use one for a course.
#ZegnatThe course may be created by a natural (you) or legal (the university) person, who could be marked up with an h-card. But the course itself is just one of possibly many courses by this person, and I feel like that makes sense as an h-entry.
#ZegnatFeel free to question and/or challenge me on that though. Not like it is set in stone. But that’s how I use h-cards.
#[jgmac1106]that makes sense, I will revise, its the syllabi that get real messy each week an h-event with due dates and with the correct markup for each APA citation
#snarfedfwiw, there's plenty of h-card usage for non-persons, eg organizations and locations
[kevinmarks] joined the channel
#[kevinmarks]A course is a collection of events? So the header could be an h-feed header
#Zegnatorganisations would be legal persons in my example, snarfed. Locations was definitely not covered by my [natural|legal] persons example though. Good catch. I am not sure I have ever used that myself.
#Loqireasonablehumanjudgmentexperimentationanddiversity has 1 karma
snarfed, sebsel and KartikPrabhu joined the channel
#sebselSo, in Microsub, you can read entries from channels, remove them and mark them as read, but you can't add them.
#sebselI know Aperture does this with a sort of Micropub-endpoint, but that's not part of this spec.
#sebselMy use case: in my experimental reader I had the function to mark things as 'read later'. That way, I could just scroll through my channels / feeds, and let the reader mark everything as read, but still let me get back to the things I found interesting.
#sebselAnother way to go would be if there's an option to get all unread items from a channel, but that's not a nice flow if you're marking as read automatically, like Together does.
swentel and [jgmac1106] joined the channel
#[jgmac1106]@zegnat well eventually I want an h-course or h-learning....but again I need a user total >1 for that to have any value
snarfed, swentel, [kevinmarks], renem and [grantcodes] joined the channel
#[grantcodes][sebsel] auto marked read is an option in together 🙂
eli_oat, snarfed, KartikPrabhu and [Sherab] joined the channel
#pstuifzandThere is a use case for keeping stuff unread, but as of yet it's not supported by the protocol.
#pstuifzandI added two ways to add items to channels, (1) adding items from the outside using a micropub like protocol, and (2) adding items from the inside, metions, for example: items that match a certain regex.
#sknebelkeeping it unread is (just don't mark it), un-marking it isn't
#pstuifzandMonocle marks items as read automatically
#sknebelyeah, but that's unrelated to the protocol
#pstuifzandthere is quite a bit of freedom to make this work
#pstuifzandmy server removes all read items from the channel
#pstuifzandand the first items in the channel is the first items you see
#pstuifzandbut it could be made so that you can move back in the channel to see the previously read items
[sebsel] joined the channel
#[sebsel][grantcodes] I saw that! But I actually really like that functionality and want to keep it like that. I just want some sort of mechanism to keep a ‘read later’-list, which is not really supported at the moment :)
[kevinmarks] joined the channel
#[kevinmarks]My podcast player has a listened to list, which is handy
#sknebel[sebsel]: well, you could create "want to read" posts on your site and subscribe to that feed ;)
tantek joined the channel
#sknebel(I'll have to figure something in that regard out too. my current reader allows to like/star posts, I'm not sure I want to convert those to full posts on my site
#LoqiA fragmention is an extension to URL syntax that links and cites a phrase within a document by using a URL fragment consisting of the phrase itself, including whitespace https://indieweb.org/fragmention
tantek, AngeloGladding and snarfed joined the channel
#[sebsel]Yeah maybe it’s more a separate service, bookmarking and reading things later. But it feels related to the Reader to me, and I would like to integrate it anyway
#sknebelyeah, certainly related. question is if it makes sense to go through your main micropub endpoint through that (which seems to be the current direction) or if another path makes more sense
#tantek.comedited /blockchain (+3089) "cluster see alsos into new subsections of a Criticism section, new descriptions section, add GDPR conflict" (view diff)
#tantek.comedited /blockchain (+70) "/* Violates GDPR */ since actual instance of blockchain death by GDPR, changing heading to be more explicit" (view diff)
#tantek.comedited /blockchain (+784) "/* Killed or being killed by GDPR */ Can blockchain’s immutability survive GDPR’s right to be forgotten?" (view diff)
#tantek.comedited /blockchain (+520) "/* Killed or being killed by GDPR */ wow this just happened today (actual blockchain death by GDPR)" (view diff)
#tantek.comedited /blockchain (+362) "/* Killed or being killed by GDPR */ better source of PICOPS death with direct links to more citations" (view diff)